Temporal variability of pyrethroid metabolite levels in bedtime, morning, and 24-hr urine samples for 50 adults in North Carolina

Marsha K. Morgana,*, Jon R. Sobusa, Dana Boyd Barrb, Carry W. Croghana, Fu-Lin Chena, Richard Walkera, Lillian Alstona, Erik Andersena, and Matthew S. Cliftona

a National Exposure Research Laboratory, US EPA, Research Triangle Park, NC, USA b Department of Environmental Health, Rollins School of Public Health, Emory University, Atlanta, GA, USA

ABSTRACT

Pyrethroid insecticides are widely used to control insects in both agricultural and residential settings worldwide. Few data are available on the temporal variability of pyrethroid metabolites in the urine of non-occupationally exposed adults. In this work, we describe the study design and sampling methodology for the Pilot Study to Estimate Human Exposures to Pyrethroids using an Exposure Reconstruction Approach (Ex-R study). Two major objectives were to quantify the concentrations of several pyrethroid metabolites in bedtime, first morning void (FMV), and 24-hr urine samples as concentration (wet weight), specific-gravity (SG) corrected, creatinine (CR) corrected, and excretion rate values for 50 Ex-R adults over a six-week monitoring period and to determine if these correction approaches for urine dilution reduced the variability of the biomarker levels. The Ex-R study was conducted at the United States Environmental Protection Agency’s Human Studies Facility in Chapel Hill, North Carolina USA and at participants’ homes within a 40-mile radius of this facility. Recruitment of participants and field activities occurred between October 2009 and May 2011. Participants, ages 19–50 years old, provided daily food, activity, and pesticide-use diaries and collected their own urine samples (bedtime, FMV, and 24-hr) during weeks 1, 2, and 6 of a six-week monitoring period. A total of 2503 urine samples were collected from the study participants. These samples were analyzed for the pyrethroid metabolites 3-phenoxybenzoic acid (3-PBA), cis/trans-3-(2,2-dichlorovinyl)-2,2-dimethyl-cyclopropane carboxylic acid (cis/trans-DCCA), and 2-methyl-3-phenylbenzoic acid (MPA) using high performance liquid chromatography/tandem mass spectrometry. Only 3-PBA was frequently detected (> 50%) in the adult urine samples. Median urinary 3-PBA levels were 0.88 ng/mL, 0.96 ng/mL-SG, 1.04 ng/mg, and 1.04 ng/min for concentration, SG-corrected, CR-corrected, and excretion rate values, respectively, across all urine samples. The results showed that median urinary 3-PBA concentrations were consistently the lowest in FMV samples (0.77 ng/mL, 0.68 ng/mL-SG, 0.68 ng/mg, and 0.58 ng/min) and the highest in 24-hr samples (0.92 ng/mL, 1.06 ng/mL-SG, 1.18 ng/mg, and 1.19 ng/min) across all four methods. Intraclass correlation coefficient (ICC) estimates for 3-PBA indicated poor reproducibility (< 0.22) for all urine sample types and methods over a day, week, and six weeks. Correcting for urine sample dilution, based on either SG, CR or urine output, introduced additional measurement variability both between- and within-individuals. These results indicate that a single measure of urinary 3-PBA was not sufficient to characterize average exposure regardless of sample type, correction method, and time frame of collection. In addition, the study results can be used to inform the design of exposure characterization strategies in relevant environmental epidemiology studies in the future.

Keywords: Pyrethroids, adults, urine, homes, biomonitoring, variability

* Corresponding author at: US EPA, 109 T.W. Alexander Dr., MDE205-04, Research Triangle Park, NC, 27709 USA. Tel.: +1 919 541 2598; fax: +1 919 541 0905. E-mail address: morgan.marsha@epa.gov (M.K. Morgan).

Abbreviations

3-PBA, 3-phenoxybenzoic acid; CDC, Centers for Disease Control and Prevention; cis/trans-DCCA, cis/trans-3-(2,2-dichlorovinyl)-2,2-dimethylcyclopropane-1-carboxylic acid; EPA, Environmental Protection Agency; Ex-R study, Pilot Study to Estimate Human Exposures to Pyrethroids using an Exposure Reconstruction Approach; HSF, Human Studies Facility; FMV, first morning void; ICC, intraclass correlation coefficient; LOQ, limit of quantitation; NHANES, National Health and Nutritional Examination Survey; NC, North Carolina; MPA, 2-methyl-3-phenylbenzoic acid; RTP, Research Triangle Park; US, United States

1. Introduction

Pyrethroids are a class of insecticides that are widely and globally used to control a

variety of insects at homes, on domestic pets, and on agricultural crops (US EPA, 2013;

Saillenfait et al., 2015). These lipophilic insecticides have replaced many of the previous

residential and some agricultural applications of the organophosphates because of their

lower volatility and lower mammalian toxicity due to faster enzymatic detoxification

(Elliot, 1976; Barr et al., 2010). However, recent research has raised concerns about

potential adverse health effects (i.e., developmental and male reproductive) occurring

from human exposure to pyrethroids at environmental levels (Saillenfait et al., 2015).

In the United States (US), at least 20 different pyrethroids are registered for

commercial use in residential or agricultural settings (US EPA, 2011). However, no

published data are currently available on the national retail sales or usage patterns of

pyrethroid insecticides in these settings (Kuivila et al., 2012; Palmquist et al., 2012; Xue

et al., 2014). Several US studies have detected a number of current-use pyrethroids (i.e.,

bifenthrin, cyfluthrin, cypermethrin, cyhalothrin, deltamethrin, and esfenvalerate) in dust,

food, and/or wipes in residential environments (Morgan et al., 2007; Julien et al., 2008;

Starr et al., 2008; Stout et al., 2009; Tulve et al., 2008; Chuang and Wilson, 2011;

Trunnelle et al., 2014). Research has suggested that dietary ingestion is likely the major

route of exposure to pyrethroids in the general US adult population (Reiderer et al., 2008;

Barr et al., 2010).

After absorption into the body, pyrethroid insecticides are metabolized rapidly and

primarily excreted in urine with an elimination half-life of less than 12-hr (Leng et al.,

1997; Kuhn et al., 1999; Ratelle et al., 2015). Several cross-sectional studies have

reported measureable concentrations of pyrethroid metabolites, including 3-

phenoxybenzoic acid (3-PBA) and cis/trans-3-(2,2-dichlorovinyl)-2,2-

dimethylcyclopropane-1-carboxylic acid (cis/trans-DCCA), in the spot urine samples of

adults recruited from the general population (referred to subsequently as non-

occupationally exposed) in the US (Berkowitz et al., 2003; Barr et al., 2010; McKelvey et

al., 2013; Young et al., 2013; Trunnelle et al., 2014; CDC, 2015; Morgan, 2015). In the

US National Health and Nutritional Examination Survey (NHANES, 2001-2002 cycle),

which was a population-based survey that included 1128 adults, ages 20–59 years, 3-

PBA, cis-DCCA, and trans-DCCA were detected in 76%, 33%, and 25% of the urine

samples, respectively (Barr et al., 2010). 3-PBA is a nonspecific urinary biomarker of at

least 18 different pyrethroid insecticides including cyhalothrin, cypermethrin,

deltamethrin, esfenvalerate, and permethrin (CDC, 2009; Barr et al., 2010). The isomers,

cis-DCCA and trans-DCCA, are also the nonspecific urinary biomarkers of cyfluthrin,

cypermethrin, and permethrin (CDC, 2009).

For spot urine measurements, a major concern is that volume-based concentrations

(e.g., ng/mL) can be affected by variable urine dilutions in humans from factors such as

fluid intake, perspiration, gender, age, and disease status (e.g., kidney and heart)

(Boeniger et al., 1993; Mage et al., 2004; Barr et al., 2005; Fortin et al., 2008). To adjust

for variable urine dilutions in spot samples, correction methods involving specific gravity

(SG) and creatinine (CR) measurements or timed excretion rates (e.g., ng/min) have been

used to normalize pesticide biomarker measurements across and within individuals

(Hwang et al., 1997; Fortin et al., 2008; Christensen et al., 2012). CR-correction of urine

sample volumes is the most widely used method in studies, and assumes that CR is

eliminated at a constant rate by the kidneys into urine in humans (Hwang et al., 1997;

Barr et al., 2005). But, this correction method may also bias study results because CR

excretion can also vary due to many factors such as age, gender, dietary intake, muscle

mass, physical activity, and diurnal patterns (Boeniger et al., 1993; Hwang et al., 1997;

Mage et al., 2004). Further complicating interpretation of these data, concentrations in

spot samples likely only reflect recent exposures because these chemicals have short

elimination half-lives (< 12-hr). Continual exposures to pyrethroid insecticides would

likely result in relatively constant urinary metabolite levels over time (e.g., day, week or

month) (NRC, 2006; Wielgomas 2013a). In contrast, episodic exposures to these

insecticides would probably cause significant fluctuations in urinary biomarker levels

over time (NRC, 2006). Little information, however, is currently known on the temporal

variability of pyrethroid metabolites in non-occupationally exposed adults. Temporal

variability is useful for understanding the patterns and magnitude of a person’s exposures

to pyrethroid insecticides over time using urinary biomonitoring. When relying on only

one spot urine measurement, this unknown variability in an individual’s biomarker level

could lead to exposure misclassification in epidemiological studies (Lassen et al., 2013).

Wielgomas (2013a) recently examined the temporal variability of 3-PBA concentrations

in the urine samples of seven, non-occupationally exposed adults (ages 24–71 years old)

over a week in an urban area of Poland in 2011. In this study, the intraclass correlation

coefficient (ICC), which quantifies reliability, were estimated for 3-PBA levels by urine

sample type (first morning void [FMV], spot, and 24-hr) as concentration, CR-corrected,

and excretion rate values. The ICCs were the highest for the CR-corrected spot urine

samples (ICC=0.846). Based on this ICC of 0.846, a single spot sample appeared

representative of average exposure over a single week, and that these adults were likely

continually exposed to similar levels of pyrethroids, likely from the same sources, in their

urban environments (Wielgomas, 2013a).

The “Pilot Study to Estimate Human Exposures to Pyrethroids using an Exposure

Reconstruction Approach” (Ex-R study) investigated longitudinal exposures (over six

weeks) of adults to pyrethroid insecticides commonly found in residential settings. The

first study objective was to assess the variability of 3-PBA, cis-DCCA, trans-DCCA, and

2-methyl-3-phenylbenzoic acid (MPA; specific metabolite of bifenthrin) in the urine

samples of 50 adults at home over a six-week monitoring period. The second study

objective was to estimate the dietary and non-dietary exposures of these adults to selected

current-use pyrethroids and to their preformed metabolites (e.g., 3-PBA) in media using

an exposure reconstruction approach. Here we report the study design and sampling

methodology for the Ex-R study, and the study results related to objective 1 above.

Specifically for objective 1, we quantified the distributions of 3-PBA, cis-DCCA, trans-

DCCA, and MPA by urine sample type (bedtime, FMV, and 24-hr) as concentration,

SG-corrected, CR-corrected, and excretion rate values over a six-week monitoring

period. We also determined if these three correction approaches for urine dilution

reduced the variability of the observed biomarker levels (3-PBA, only). In addition, we

calculated the number of urine samples needed to adequately assess the exposures of Ex-

R adults to pyrethroids insecticides over a day, week, and six weeks.

2. Materials and methods

The design strategy, recruitment of participants, field sampling procedures, and

analytical methodology (urine, only) for the Ex-R study are discussed below.

2.1. Study cohort

The Ex-R study investigated the longitudinal exposures of 50 adults to pyrethroid

insecticides at their residences over a six-week monitoring period. The study was

conducted at the US Environmental Protection Agency’s (EPA’s) Human Studies Facility

(HSF) in Chapel Hill, North Carolina (NC) and at participants’ homes within a 40-mile

radius of the HSF. From October 2009 to March 2011, adults were recruited by an on-site

US EPA contractor (WESTAT) at the HSF via their in-house database of volunteers or by

word of mouth (i.e., US EPA researchers or prior participants). Interested individuals

were asked several questions (e.g., “Are you exposed at work to products that contain

pyrethroid insecticides [yes or no]?)” over the phone by a WESTAT technician to

determine their eligibility to participate. To be eligible, a participant had to be a healthy,

non-pregnant adult (18-50 years old), had no pre-existing medical conditions (e.g.,

kidney disease) affecting urine output, was currently living in a residential home or

apartment, had no occupational exposures to pyrethroids, could provide their own

transportation to and from the HSF, and could speak, read, and use English fluently. A

total of 56 adults who contacted WESTAT were eligible to participate in the study; only

two of these adults declined to provide consent to participate (e.g., work conflict). Before

beginning the study, the adult participants were trained individually (~2 hrs) by US EPA

researchers at the HSF on the procedures for filling out diaries (activity, food, and

pesticide-use), collecting their own samples (urine, solid food, drinking water, surface

wipes, and vacuum dust), and storing the samples in portable thermoelectric coolers until

collected by study personnel.

2.2 Human subjects protection

The study protocol and procedures to acquire informed consent from the adult

participants followed EPA policies and the guidelines set forth by the Scientific and

Ethical Approaches for Observational Exposure Studies report (US EPA, 2008) and were

reviewed and approved by the US EPA’s Human Subjects Research Review Official and

by the University of North Carolina’s Institutional Review Board (study number 09-

0741). Adult volunteers read and signed informed consent documents prior to beginning

the study.

2.3. Field sampling

Field activities were performed between November 2009 and May 2011. A total of 11

consecutive, six-week monitoring periods were performed in this study. During each

monitoring period, up to six adult participants filled out diaries and collected their own

samples during weeks 1, 2 and 6. Fig. 1 presents the sampling schedule that the

participants followed for each sampling week. Each sampling week began in the morning

of day 1 (Sunday) and ended in the morning of day 6 (Friday). There were two sampling

intervals, days 1-3 (Sunday to Tuesday) and days 4-6 (Wednesday to Friday), per

sampling week.

The participants used portable thermoelectric coolers (13.5"D x 12" W x 14.5"H

Princess International or Vinotemp®) to store their samples and diaries (in an

outside pocket) for each sampling interval. Each cooler contained a notebook with

sampling instructions, calendar and schedule, a notebook containing food, activity, and/or

pesticide-use diaries, sampling equipment and containers, gloves, ballpoint pens, and an

electric power adaptor (Koolatron® Model AC-15). Barcodes were attached to all diaries

and sampling containers. Temperature data recorders (Easy Log EL-USB-LITE or EL-

USB-1, Lascar Electronic, Ltd) were placed inside each cooler to record internal

temperatures. The participants kept the coolers at reduced temperatures by plugging them

into their vehicle (using a car adaptor) or an electric wall outlet at home or other location

(e.g., work).

The participants were provided with two coolers for the first sampling interval (days

1-3) on the Friday before each sampling week. These two coolers were returned by the

participants to the HSF between 8:00 am–11:00 am on day 3 of each sampling week.

After arriving at the HSF, US EPA researchers checked-in all study items in the two

coolers with each participant. Afterwards, the participants were provided with two new

coolers for the next sampling interval (days 4-6) of each sampling week. These two

coolers were returned to the HSF between 8:00 am–11:00 am on day 6 of each sampling

week, and all study items were checked-in by US EPA researchers. After each sampling

interval, the used coolers containing collected study items with blue ice were then

transported (~20 miles) by US EPA researchers in a van to the US EPA laboratory in

Research Triangle Park (RTP), NC. The participants came to the HSF a total of eight

times, including the training session, during the six-week monitoring period.

2.4. Collection of diaries

a Period 1 (P1) = 4:00 am–11:00 am; Period 2 (P2) = 11:00 am–5:00 pm; Period 3 (P3) = 5:00 pm–4:00 am b Bedtime void (B), first morning void (M), and 24-hr sample c Individual urine voids were collected over the 24-hr period and analyzed separately d Vacuum dust sample was collected only on day 4 of week 6 e Drinking water sample was collected only on day 3 of week 6 f Study items were dropped off at the HSF between 8:00 am – 11:00 am

Fig. 1. Field sampling activities that the Ex-R adult participants performed during each sampling week (1, 2, or 6).

The participants filled out three different types of diaries (food, activity, and pesticide-

use) during sampling weeks 1, 2, and 6 (Fig. 1). For the food and activity diaries,

information was collected during three consecutive time periods on days 1-2 and days 4-5

of each sampling week. These time periods were period 1 (4:00 am -11:00 am), period 2

(11:00 am - 5:00 pm), and period 3 (5:00 pm to 4:00 am). In the food diary, the

participants recorded the types and amount of foods and beverages they consumed during

each time period. In the activity diary, the participants recorded their primary location

(i.e., inside home, outside home, away from home, or transit) and primary activity (i.e.,

sleeping, low activity [e.g., sitting], medium activity [e.g., walking] or high activity [e.g.,

running]) in 30-minute time intervals for each time period. They also recorded when they

ate, drank, or urinated during these 30-minute time intervals. For the pesticide-use diary,

information was collected on day 5 of each sampling week. In this diary, the participants

recorded the types of insecticides that were used at their home over a period of 30 days

prior to and during the six-week monitoring period. They also recorded their pesticide

usage patterns including type, location, frequency, and day of insecticide application

during each sampling week.

2.5. Collection of environmental samples

Environmental samples including solid food, drinking water, surface wipes, and

vacuum dust were collected by the adult participants at their residences during sampling

weeks 1, 2, and/or 6 (Fig. 1). Detailed information on the collection of these media by the

participants can be found in the Supplemental Information section.

2.6. Collection of urine samples

Spot urine voids (bedtime and FMV) and 24-hr urine samples were collected by the

participants during each sampling week (Fig. 1). Bedtime voids were collected before

going to sleep on days 1 and 4 of each sampling week. FMV’s were collected after

waking up on days 2 and 5 of each sampling week. For the 24-hr samples, individual

urine voids for a person were collected starting after the FMV on day 2 or 5 until

collection of the FMV the next day (e.g., day 3 or 6) (Fig. 1). Up to 11 individual voids

were collected by a participant over a 24-hr sampling period.

All urine voids were collected by the participants in separate 1 L polypropylene

containers. The entire amount of urine was collected for each type of urine void. The

time of sample collection for each void was recorded on a label affixed to the lid of the

container. At the HSF, US EPA researchers recorded the volume of each urine void using

a transparent, graduated plastic cylinder that was placed over the container. A total of

2503 urine voids (includes bedtime, FMV, and 24-h) were collected by the participants

during the study.

At the US EPA laboratory, up to eight, 8 mL aliquots of urine from each void were

placed into separate 10 mL cryogenic vials with lids. All aliquots were stored in

laboratory freezers (-80oC) until analysis.

2.7 Analysis of urine samples

In 2010–2011, US EPA laboratory technicians shipped 2503 urine aliquots (8 mL

each) in coolers containing dry ice overnight by UPS to Emory University’s Rollins

School of Public Health laboratory in Atlanta, Georgia. The urine aliquots were stored in

-80oC freezers until analyzed (2013–2014). For the 24-hr urine samples, each void was

analyzed separately. All urine samples were extracted and levels of 3-PBA, cis-DCCA,

trans-DCCA, and MPA were measured using a slight modification of the method by

Baker et al. (2004). Briefly, 0.5-mL aliquot of urine from each sample was hydrolyzed

with β-glucuronidase/sulfatase for ~16 hrs. Next, the target analytes were extracted using

solid phase extraction (3cc-60 mg OASIS HLB cartridge) and then evaporated to dryness

using a TurboVap LV Evaporator set at 45oC and 15 psi N2. Finally, the extract was

reconstituted with 100 µL of a 30% methanol/70% deionized water solution prior to

chemical analysis. The urine extracts were quantified for the target pyrethroid metabolites

using a high performance liquid chromatography/tandem mass spectrometer (Agilent

Tech, Waldbronn, Germany) equipped with a negative mode jet-stream electrospray

chemical ionization interface. Separation was performed using a Betasil C18, 100 x 2.1

mm - 3 µm particle size, column (Thermo Scientific, Waltham, MA, USA) with a

gradient elution using 0.1% acetic acid in deionized water and 0.1% acetic acid in

methanol. The column temperature was maintained at 45oC, and the total run time was

seven minutes. The estimated limits of quantitation (LOQ) were 0.25 ng/mL for 3-PBA,

cis-DCCA, and trans-DCCA and 0.5 ng/mL for MPA.

All urine samples were also analyzed for SG and CR levels. At the US EPA laboratory

in 2012, another identical set of urine aliquots (8 mL each) was thawed overnight in a

refrigerator (~4oC). Each aliquot of urine was vortex mixed (~5 seconds) and then

transferred into a 50 mL vial and sonicated for 15 minutes. A 50 µL aliquot of urine was

removed from each 50 mL vial and analyzed for CR levels using a spectrophotometer

(Biotek® ELx800 or Elx808ui), equipped with a 96-well plate at a wavelength at 490 nm

using a modified version of the Jaffé method (Andersen et al., 2014). SG was also

measured by dipping the tip of a hand-held refractometer (Atago® model no. 3741) into

each 50 mL vial of remaining urine (~7.9 mL) using water as a calibrant.

2.8 Quality assurance and quality control

A total of 150 field blanks and 150 field spikes were prepared prior to the onset of

sampling activities by laboratory technicians at the Centers for Disease Control and

Prevention (CDC) in Atlanta, Georgia in 2009. These quality control (QC) samples

consisted of pooled urine samples that were collected from non-occupationally exposed

adult volunteers at CDC. The pooled urine samples were prescreened for 3-PBA, cis-

DCCA, and trans-DCCA and diluted, when necessary, to reduce background levels. Each

field blank or field spike consisted of 8 mL of urine in a 10-mL polypropylene cryovial.

For the field spikes, 3-PBA, cis-DCCA, and trans-DCCA standards (25 ng/mL) were

individually added to each cryovial; the MPA standard was not obtained until after

preparation of the field spikes at CDC. These QC samples were shipped in coolers on dry

ice overnight by Fed-Ex from the CDC laboratory to the US EPA laboratory in 2009 and

kept frozen in freezers (-80oC) until field sampling.

A single field blank and field spike were assigned to each study participant during

days 4-6 of each sampling week. These QC samples were thawed overnight in an US

EPA refrigerator and then placed unopened in a (plugged-in) portable thermoelectric

cooler in a study room at the HSF. After each sampling interval, the field blanks and field

spikes were transported in separate coolers along with study sample coolers to the US

EPA laboratory and kept frozen (-80oC) in freezers.

In 2010–2011, 150 field blanks, 150 field spikes, and 150 duplicate urine samples

were shipped in coolers on dry ice by UPS from the US EPA laboratory to the Emory

University laboratory in Atlanta, GA. These QC samples were analyzed for the target

pyrethroid metabolites following the same analytical procedures described in section 2.7

above. For the field blanks and field spikes, a value of LOQ¿√2 was first imputed for all

target analyte measurements that were below the LOQ (Verbovsek, 2011). Precision

estimates, expressed as percent relative standard deviations (RSDs), were then calculated

for the field blanks and field spikes. Measures of field blanks were all below the LOQ for

cis-DCCA, trans-DCCA, and MPA. However, all field blanks had low 3-PBA

background levels (likely stemming from true background levels in the pooled urine

samples, and not sample contamination), and yielded an RSD estimate of 2.5%. For the

field spikes, estimates of percent relative standard deviation across the samples were

between 6% and 7% for 3-PBA, cis-DCCA, and trans-DCCA. Precision across duplicate

urine samples was evaluated using relative percent difference estimates at both the

metabolite and sample levels. Not including zero values (where the initial and replicate

measures were identical – generally a consequence of single value imputation for

measurements <LOQ), median relative percent differences across duplicate sample

measures were estimated to be 17% for 3-PBA (based on 129 samples), 26% for cis-

DCCA (based on 64 samples), 18% for trans-DCCA (based on 80 samples), and 41% for

MPA (based on 4 samples).

In addition to field QCs, laboratory QC samples (~15% of the samples tested) were

used to evaluate the quality of each analytic run. A minimum of eight calibrants, 2 QC

materials, 2 blank samples, and 1 replicate sample were analyzed concurrently with 30-40

unknown samples in each analytic run. QC samples had to be within 20% of the spiked

concentration for the run to be considered valid. Replicate samples were not used to

determine run validity but were rather used as a quality check. Replicate samples were

generally within 20% of their respective duplicate.

2.9. Statistical analysis

All statistical analyses were performed using SAS version 9.4 (SAS Institute, Cary,

NC, USA). For each pyrethroid metabolite, all urine data values below the LOQ were

assigned the value of LOQ¿√2 (Verbovsek, 2011). Descriptive statistics including

geometric mean, range, and selected percentiles (25th, 50th, 75th, 95th, and 99th) were

calculated for the concentrations of 3-PBA, cis-DCCA, and trans-DCCA by urine sample

type (bedtime, FMV, 24-hr, and overall) as concentration (ng/mL), SG-corrected (ng/mL-

SG), CR-corrected (ng/mg), and excretion rate (ng/min) values. These descriptive

statistics were not calculated for MPA as it was detected in only 2% of the urine samples.

SG-corrected values were computed using the following equation (Morgan et al., 2008):

SG-corrected value (ng/mL-SG) = urine concentration (ng/mL) x (SGtarget - 1.000)/(SGurine

sample - 1.000), where SGtarget was selected to be the median SG value (1.014) across all

study measurements. SG values ranged from 1.0018–1.0372 ng/mL-SG. CR-corrected

values were calculated using the following equation (Morgan et al., 2008): CR-corrected

values (ng/mg) = 100 mL/dL x urine concentration (ng/mL)/CR concentration (mg/dL).

CR concentration values ranged from 7.0 – 442 mg/dL. The metabolite excretion rate was

calculated as follows: Excretion rate (ng/min) = urine concentration (ng/mL) x volume of

urine void (mL)/time since previous urine void (min). Urine void volumes ranged from 5

mL to 1000 mL (limited by container size of 1 L).

Pairwise associations between urinary 3-PBA, cis-DCCA, and trans-DCCA levels

were evaluated using Spearman correlation coefficients. Since repeated measures from

individuals are not independent, the results of these analyses were used only to screen for

general trends between the urinary metabolites.

Urinary 3-PBA was the only frequently detected analyte (>50%), and therefore the

only analyte on which parametric analyses were performed. Based on the Shapiro-Wilks

normality test, the distribution of urinary 3-PBA was found to be log-normal, and

therefore all urine values were log-transformed (ln) for further statistical analyses. For the

frequently detected 3-PBA, one-way random effects models (Proc Mixed) were used to

estimate within- and between-person variance components for ln-transformed 3-PBA.

These estimates were then used to estimate ICCs and 95% fold-ranges (R0.95) according to

Rappaport and Kupper, 2008. The ICC represents the ratio of between-person variance

to total variance. ICC values can range from 0 to 1; values nearer to 0 signify low

reliability, and values nearer to 1 signify high reliability (Fleiss, 1985). ICCs values can

be used to indicate poor reproducibility (<0.4), good reproducibility (0.4 to 0.75) or

excellent reproducibility (> 0.75) of a spot urine measurement (Rosner, 2006).

Traditionally, an ICC value of 0.80 or greater indicates that a random spot urine

measurement would accurately represent the average biomarker value over a defined

period of time (Fleiss, 1985). The between-person fold-range (bR0.95) represents the fold-

range containing the middle 95% of individual-specific mean biomarker levels. The

within-person fold-range (wR0.95) represents the fold-range containing the middle 95% of

biomarker levels for any individual. ICC and fold-range estimates for ln(urinary 3-PBA

levels) were calculated separately by urine sample type as concentration, SG-corrected,

CR-corrected, and excretion rate measures corresponding to a day, week, and six-weeks.

Lastly, we estimated the number of random spot urine voids per adult that would be

required to provide a reliable biomarker estimate (ICC = 0.80) by sample type and

method for each time period using the following equation (Fleiss, 1985): m = (pr,m(1 –

pr)) / (pr(1 – pr,m)). In this equation, m is the number of random spot urine measurements

per adult required to correctly rank individuals within a cohort, pr,m is the specified

reliability of the mean (ICC= 0.80), and pr is the ICC value for each urine sample type

and method.

3. Results

Table 1 presents the physical characteristics including gender, age, weight, height, and

race of the 50 adult participants in the Ex-R study. There were a total of 30 females and

20 males, and their ages ranged from 19-50 years old. All of the participants completed

the six-week monitoring period, except for one female who dropped out of the study due

to a family emergency during the last sampling interval (days 4-6) of week six (Fig. 1.).

Four additional participants were dropped from the study before completing the six-week

monitoring period because they did not properly follow sampling procedures (e.g.,

missed collection of several solid food and/or urine samples).

During sampling weeks 1, 2, and 6 a total of 12%, 14%, and 12% of the participants,

respectively, reported that a product containing a pyrethroid was applied to kill insect

pests at their residences. In addition, during non-sampling weeks, 10% of the participants

reported that pyrethroid products had been used for insect control at their homes.

Targeted

insects included ants, cockroaches, crickets, fleas, mosquitoes, silverfish, spiders, and/or

ticks. Pyrethroids contained in these applied products at the homes were allethrin,

bifenthrin, cyfluthrin, cyhalothrin, cypermethrin, cyphenothrin, deltamethrin,

esfenvalerate, imiprothrin, metofluthrin, permethrin, phenothrin, prallethrin and

tralomethrin. Based on this information, at least 70% of the applied pyrethroid

insecticides at the participants’ homes (within a month and during the six-week

monitoring period) can be metabolized to form 3-PBA, cis-DCCA, trans-DCCA, and/or

MPA in the body (Fig. 2).

The participants collected on average 8 urine voids (range = 3–11) per sampling day.

The pyrethroid metabolites, 3-PBA, cis-DCCA, trans-DCCA, and MPA were detected in

74%, 32%, 39%, and 2% of the urine samples, respectively. Table 2 provides the

descriptive statistics for the urinary levels of 3-PBA, cis-DCCA, and trans-DCCA by

urine sample type and method. In all of the urine samples at the 75th and 95th percentiles,

urinary levels of 3-PBA were consistently higher compared to urinary levels of cis-

DCCA and trans-DCCA for all methods. For 3-PBA (the only metabolite with

measurements >LOQ at the 50th percentile), median urinary levels were consistently the

lowest in FMV samples

Table 1Physical characteristics of the adult participants in the Ex-R study.

Characteristic All Females MalesSex (number)Age (years) Weight (kilogram) Height (cm) Raceb

Non-Hispanic White Non-Hispanic Black Hispanic Asian Native American

5033.4±9.1 (19-50)a

82.2±18.9 (48.1-130)170±8.5 (149-191)

25 (56%)c

11 (25%)6 (13%)2 (4%)1 (2%)

3034.8±9.4 (21-50)

75.6±17.9 (48.1-130)165±5.8 (149-191)

14 (31%)9 (20%)3 (7%)1 (2%)0 (0%)

2031.3±8.5 (19-48)

92.0±16.2 (57.6-130)178±5.3 (166-188)

11 (25%)2 (4%)3 (7%)1 (2%)1 (2%)

a Mean ± standard deviation and rangeb Race data were collected ~ 4 months after the study was completed. These data were collected from 45 out of 50 of the participantsc Number and percentage of total

a 3-phenoxybenzoic acid (3-PBA), cis/trans-3-(2,2-dichlorovinyl)-2,2-dimethylcyclopropane-1-carboxylic acid (cis-DCCA), and 2-methyl-3-phenylbenzoic acid (MPA)

Fig. 2. Urinary metabolites measured in Ex-R adults and their corresponding parent pyrethroid(s) used at study homes one month before and during the six-week monitoring period a

Table 2Urinary 3-PBA, cis-DCCA, and trans-DCCA levels in 50 Ex-R adults over a six-week monitoring periodab

Type of Urine Sample Nc %d Unite GMh GSDi Percentiles Max.25th 50th 75th 95th 99th Bedtime 3-PBA

cis-DCCA

trans-DCCA

291288288285291288288285290287287284

77

27

33

ng/mLng/mL-SGf

ng/mgf

ng/ming

ng/mLng/ml-SGng/mgng/minng/mLng/ml-SGng/mgng/min

0.960.880.940.90----j

----------------------------

4.024.334.675.13--------------------------------

0.350.310.290.25

<<<<<<<<

0.820.810.800.83

<<<<<<<<

2.152.382.522.870.380.500.590.590.740.670.830.84

16.314.917.219.33.252.343.513.475.904.876.697.33

37.631.233.346.411.49.008.8513.726.722.427.728.9

101.943.950.646.817.014.917.022.836.627.436.431.9

FMV 3-PBA

cis-DCCA

trans-DCCA

296295295293298297297295297296296294

70

32

38

ng/mLng/mL-SGng/mgng/minng/mLng/mL-SGng/mgng/minng/mLng/mL-SGng/mgng/min

0.830.760.740.62

<<<<<<<<

3.994.134.294.25

<<<<<<<<

<<<<<<<<<<<<

0.770.680.680.58

<<<<<<<<

2.342.162.051.640.500.450.460.370.990.840.860.62

7.548.359.787.692.562.232.712.095.624.714.544.24

92.356.145.654.815.313.813.512.417.814.914.615.3

180.6115.1116.4134.639.424.019.523.467.140.840.239.8

24-hr 3-PBA

cis-DCCA

trans-DCCA

187718561856186418881866186618751887186518651874

74

33

41

ng/mLng/mL-SGng/mgng/minng/mLng/mL-SGng/mgng/minng/mLng/mL-SGng/mgng/min

0.981.171.301.32

<<<<<<<<

4.294.805.285.38

<<<<<<<<

<<<<<<<<<<<<

0.921.061.181.19

<<<<<<<<

2.393.243.753.950.540.710.900.951.031.201.401.53

12.318.724.323.73.284.455.625.337.549.7312.312.3

64.780.4115.5106.118.225.030.335.042.339.648.262.5

3154.91983.91784.52035.4702.5441.8397.3453.2197.7291.4516.1563.8

All 3-PBA

cis-DCCA

trans-DCCA

247224462446244924852458245824622482245524552459

74

32

39

ng/mLng/mL-SGng/mgng/minng/mLng/mL-SGng/mgng/minng/mLng/mL-SGng/mgng/min

0.961.071.171.15

<<<<<<<<

4.224.705.135.31

<<<<<<<<

<<<<<<<<<<<<

0.880.961.041.04

<<<<<<<<

2.332.883.303.470.500.640.810.840.991.081.191.29

12.216.520.621.33.163.844.854.617.158.6610.39.64

56.472.693.293.916.920.224.930.732.935.443.252.3

3154.91983.91784.52035.4702.5441.8397.3453.2197.7291.4516.1563.8

a 3-Phenoxybenzoic acid (3-PBA), cis-3-(2,2-dichlorovinyl)-2,2-dimethylcyclopropane-1-carboxylic acid (cis-DCCA), and trans-3-(2,2-dichlorovinyl)-2,2-dimethylcyclopropane carboxylic acid (trans-DCCA).b 2-methyl-3-phenylbenzoic acid (MPA) values are not reported because it was detected in 2% of the urine samples.c Number of urine samples (note: not all 2503 samples could be analyzed due to limited sample volume) d Percentage of urine samples at or above the limit of quantitation (LOQ)e Data are provided as concentration (ng/mL), specific gravity-corrected (ng/mL-SG), creatinine-corrected (ng/mg), and excretion rate (ng/min) valuesf 26 urine samples were excluded because of missing creatinine and specific gravity data due to laboratory errorg 17 urine samples were excluded due to missing or partial voids provided by the adult participantsh Geometric meani Geometric standard deviation (unitless)j The geometric mean and geometric standard deviation are not reported for pyrethroid metabolites that were detected in less than 50% of the urine samples.k Below the LOQ (<0.25 ng/mL) for a pyrethroid metabolite in the urine samples.

(0.77 ng/mL, 0.68 ng/mL-SG, 0.68 ng/mg, and 0.58 ng/min) and the highest in 24-hr

samples (0.92 ng/mL, 1.06 ng/mL-SG, 1.18 ng/mg, and 1.19 ng/min) across all four

methods.

Based on Spearman correlation coefficients, cis-DCCA and trans-DCCA levels

(ng/mL) were highly correlated (r = 0.86, p<0.0001) with each other in the adult urine

samples. In addition, levels of cis-DCCA (r = 0.64, p<0.0001) and trans-DCCA (r = 0.69,

p<0.0001) were strongly correlated with 3-PBA levels.

Table 3 presents ICC and fold-range estimates for repeated 3-PBA measurements by

urine sample type and method for the Ex-R adults over a day, week, and six weeks. ICCs

indicated poor reproducibility (< 0.22) for all urine sample types and methods over each

time period. In general, ICCs for SG-corrected, CR-corrected, and excretion rate values

were only slightly higher than those based on unadjusted concentration values

(suggesting a larger proportion of between-person variance to total variance). All of the

ICCs indicated a low level of reliability of repeated 3-PBA measurements in spot urine

samples for each method over a day, week, or six weeks. The number of random spot

urine samples required to provide a reliable 3-PBA biomarker estimate for an individual

ranged from 15 to 800 samples depending on the sample type and method used. CR-

corrected FMV urine samples had the lowest number of urine samples (n=15) needed to

provide a reliable 3-PBA biomarker estimate (over a week). Regardless of the sample

type and method, an unreasonably high number of spot urine samples would be required

to provide a reliable estimate of the average 3-PBA concentration for the Ex-R

participants over a day or longer.

Table 3Variance components and other related statistics for 3-PBA levels in adults by urine sample type and method over a day, week, and six-weeksa

Urine sample type

ng/mL ng/mL-SG ng/mg ng/min

bR0.95 b

wR0.95 c ICCd me

bR0.95 wR0.95 ICC m bR0.95 wR0.95 ICC m bR0.95 wR0.95 ICC m

One Day24-hrf 2.1 182 0.02 204 4.9 243 0.08 47 9.6 325 0.13 26 7.9 341 0.11 32One Week Bedtime 1.0 360 0.00 ---h 1.0 441 0.00 --- 2.8 500 0.03 144 6.8 514 0.09 43FMVg 4.0 210 0.06 60 9.3 175 0.16 22 13.9 163 0.21 15 10.6 220 0.16 2124-hr 3.7 214 0.06 66 6.1 306 0.09 40 9.8 393 0.13 27 8.4 449 0.11 33Six WeeksBedtime 1.5 232 0.01 800 3.6 271 0.05 77 5.9 322 0.09 42 6.5 460 0.09 43FMV 3.0 204 0.04 95 5.9 196 0.10 35 7.1 212 0.12 30 5.2 228 0.08 4324-hr 2.7 278 0.03 128 7.6 333 0.11 33 12.5 412 0.15 23 9.6 490 0.12 30a Urine data (log-transformed) are provided as concentration (ng/mL), specific gravity-corrected (ng/mL-SG), creatinine-corrected (ng/mg), and excretion rate (ng/min) values.b Between-person fold range c Within-person fold ranged Intraclass correlation coefficiente Number of random spot urine samples per adult likely needed to have a reliable biomarker estimate (ICC = 0.80) (Fleiss, 1985).f Results are limited to the first 24-hr interval of week one as the completion rates were the highest for this interval (88%) compared to all intervals (6 total).g First morning void h The between-person variance was zero resulting in an ICC of 0.00, therefore, no sample size could be estimated.

Between- and within-person fold-range estimates in Table 3 support ICC estimates

and further indicate that 3-PBA measurements were far more variable within individuals

than between individuals. In addition to producing the smallest ICC estimates, 3-PBA

concentration values also produced the smallest between- and within-person fold-ranges.

Results for 3-PBA concentrations suggest that the central 95% of individual-specific

mean levels were observed across a 1 to 4–fold range, whereas the central 95% of

individual measures for a given person were observed across a 182 to 360-fold range,

depending on sample type and evaluation period. In contrast to the concentration values,

3-PBA excretion rate values had the highest within-person fold ranges, and the second

highest between-person fold ranges. The highest between-person fold ranges were

observed for CR-corrected values. As shown in Table 3, the central 95% of mean CR-

corrected levels were observed across a 2.8 to 13.9-fold range, and the central 95% of

individual excretion rate measures for a given person were observed across a 220 to 514–

fold range. These results in Table 3 indicate that correcting for urine sample dilution,

based on either SG, CR, or urine output, actually introduced measurement variability,

both between- and within-individuals, when evaluating urinary 3-PBA levels.

4. Discussion

The Ex-R study was designed to be a participatory-based, exposure measurements

study that investigated the longitudinal exposures of 50 adults to pyrethroid insecticides

and to their preformed metabolites in media at their homes in NC in 2009-2011. We

found that participatory-based sampling was quite an effective way to engage and

actively involve the adult volunteers in the Ex-R study. The 50 participants filled out a

total of 299 (99%) food, 299 (99%) activity, and 149 (99%) pesticide-use diaries. They

also collected a total of 783 (99%) solid food, 50 (100%) drinking water, 149 (99%)

surface wipe, 48 (96%) vacuum dust, and 2502 (98%) urine samples during the study.

The 100% and 98% completeness for the recruitment of 50 participants and field

samples/diaries surpassed our study completeness goals of >85% and >95%, respectively.

In addition, despite the subject burden (~10-hr per person a week), participatory-based

sampling resulted in few problems in the field. The most frequently encountered

problems by study participants included malfunctioning temperature data loggers due to

20

low batteries, moisture occurring in the temperature data loggers due to cooler

condensation, loose electric power adaptor/cooler connections, broken cooler adaptors,

and broken cooler zippers. We quickly corrected the temperature data logger problems by

purchasing a new model (EL-USB-1) that had a longer battery life and by placing the

new loggers into 10 mL cryogenic vials with lids to eliminate condensation when inside

the coolers. The loose power adaptor problem was easily fixed by using an inexpensive

($5) Velco strap (Speedtech International Inc., Racine, Wisconsin) that snuggly

connected the power adaptor with the cooler plug.

We believe this is the first study in the literature to report the urinary levels of MPA,

the specific metabolite of bifenthrin, in non-occupationally exposed adults in the US.

Bifenthrin is currently registered by the US EPA for use in both agricultural and

residential settings (NPIC, 2011). In our study, conducted in 2009–2011, MPA was

detected in less than 3% of the Ex-R adult urine samples (maximum=8.9 ng/mL). This

information suggests that the majority of the participants were probably not exposed to

measureable levels of bifenthrin residues in food or residential media during the six-week

monitoring period in NC. These results are supported by a recent study (Tao et al., 2013)

showing that MPA was not detected (<0.04 ng/mL) in the spot urine samples of 10 US

children, ages 3-11 years old, in Massachusetts, USA.

Our biomonitoring results, however, did show that 3-PBA was frequently detected

(74%) in the Ex-R adult urine samples. This is in agreement with other published US

studies that have also found 3-PBA detected often (>55%) in the urine of non-

occupationally exposed adults (Berkowitz et al., 2003; Barr et al., 2010; McKelvey et al.,

2013; Young et al., 2013; Trunnelle et al., 2014; Morgan, 2015). Interestingly, the

geometric mean 3-PBA levels (0.96 ng/mL) in Ex-R participants were twice as high

compared to the geometric mean 3-PBA levels (0.42 ng/mL) in 2009-2010 NHANES

adults (CDC, 2015). Trunnelle et al. (2014) also recently reported two times higher

median urinary 3-PBA concentrations (0.82 ng/mL) in 90 Californian adults in 2007-

2009 in comparison to the 2009–2010 NHANES adults (0.39 ng/mL). This information

suggests that there is probably geographical differences in US adult exposures to

pyrethroid insecticides that are metabolized to form 3-PBA in the body. Studies

conducted outside the contiguous US have also reported 3-PBA being frequently detected

21

(>70%) in the urine of the general adult population, except for one in Puerto Rico (46%)

(Table 4). In these

Table 4Comparison of urinary 3-PBA levels in non-occupationally exposed adults worldwide.

Country Study Year Na Age %b GMc 95th d Referenceng/mLCanada 2005 120 18–64 99 ----e 4.23 Fortin et al., 2008Caribbeanf 2008-2011 297 20–39 100 0.54 3.51 Dewailly et al., 2014China 2009-2010 1149 17–45 99 0.97 5.39 Qi et al., 2012Germany 2000g 46 17–61 70 ---- 0.67 Schettgen et al., 2002Japan 2005 448 39–85 98 0.29 1.96 Ueyama et al., 2009Poland 2010-2011 132 5–77g 80 0.26 1.15 Wielgomas et al., 2013bPuerto Rico 2010-2012 54 18–40 46 0.20 2.30 Lewis et al., 2014USA (NHANES)h 2007-2008

2009-201011101296

20–5920–59

----

0.430.42

6.656.95

CDC, 2015CDC, 2015

This study 2009-2011 50 19–50 74 0.96 12.2 ----ng/mgChina 2009-2010 1149 17–45 99 1.53 8.11 Qi et al., 2012Japan 2005 448 39–85 98 0.40 2.33 Ueyama et al., 2009Poland 2010-2011 132 5–77 80 0.22 0.95 Wielgomas et al., 2013bUSA (NHANES) 2007-2008

2009-201011101296

20–5920–59

----

0.440.42

6.294.72

CDC, 2015CDC, 2015

This study 2009-2011 50 19–50 74 1.17 20.6 ----a Number of study participantsb Percentage of urine samples with detectable levels of 3-PBAc Geometric meand Percentilee Not providedf Includes ten different countries g Eight of the participants were children below the age of 14 years oldh NHANES adults between 20-59 years old, only

available studies, geometric mean urinary 3-PBA concentrations were the lowest for

Puerto Rico, Poland, and Japan (<0.30 ng/mL) and the highest for China (0.97 ng/mL).

Our study results (GM=0.96 ng/mL) were similar to the reported geometric mean 3-PBA

levels of 0.97 ng/mL reported in 1149 women from China in 2009-2010 (Qi et al., 2012).

This likely

reflects the increased usage of pyrethroid insecticides in residential settings and some

agricultural settings in the US and China (Barr et al., 2010; Qi et al., 2012).

We found that urinary cis-DCCA and trans-DCCA levels were highly correlated with

each other (r = 0.86, p<0.0001) and with urinary 3-PBA levels (cis-DCCA: r = 0.64,

p<0.0001; trans-DCCA: r = 0.69, p<0.0001) in the Ex-R adult urine samples. Similar

results were observed by Barr et al. (2010) showing strong correlations occurring

22

between the concentrations of these two isomers (r= 0.89, p<0.001) and 3-PBA (r=0.77,

p=0.02) in the urine samples of 1999–2002 NHANES adults. Only the current-use

pyrethroids, permethrin and cypermethrin, are known to breakdown to form both 3-PBA

and DCCA in the body (Tao et al., 2013; CDC, 2009). The strong correlations occurring

between these urinary biomarkers suggest that the Ex-R adults were likely primarily

exposed to these two pyrethroids in their residential environments (Barr et al., 2010).

This is in agreement with recent research by Xue et al. (2014) indicating that permethrin

and cypermethrin likely contributed to about 80% of the observed urinary levels of 3-

PBA and DCCA in the general US population.

Our ICC results indicate poor reproducibility (<0.22) of repeated 3-PBA

measurements in the spot urine samples of the Ex-R adults regardless of the sample type

(bedtime, FMV, or 24-hr) and method (non-corrected or corrected) over a day, week, and

six weeks. This ICC information suggests that a random spot urine sample did not

provide a reliable estimate of the average 3-PBA concentration for the adults in this

study. In addition, we found regardless of sample type or method used an unreasonably

high number of random spot urine samples (15-800) would likely be required per person

to provide a reliable biomarker estimate over a day or longer--which is unrealistic for

most exposure and epidemiological studies due to budget constraints and participant

burden. Our results also imply that the Ex-R adults were likely not continually exposed to

low levels of pyrethroid residues in consumed foods and/or other residential media (i.e.,

dust and surfaces) over the six-week monitoring period. Our study results are in contrast

to Wielgomas (2013a) that reported much higher ICCs for 3-PBA measurements in FMV,

spot, and 24-hr urine samples by concentration, CR-corrected, and excretion rate values

for seven non-occupationally exposed adults over a week in an urban area of Poland in

2011. In that study, the highest ICCs occurred in the CR-corrected spot (ICC=0.846) and

the CR-corrected 24-hr urine samples (ICC=0.796). Based on these ICCs, the authors

concluded that a random spot urine sample would adequately represent the average 3-

PBA biomarker concentration for an individual over a week, and that the Polish adults

were likely continually exposed to similar levels of pyrethroid residues in their urban

environments. This above research suggests that the temporal usage patterns and/or

23

exposure patterns of adults to pyrethroids may be quite different in Poland compared to

the US and more research is needed worldwide.

In our current study, we found that comparisons of biomarker measurement variance

and ICC estimates across concentration, SG-corrected, CR-corrected, and excretion rate

measures yielded unexpected results. It is generally assumed that, for urinary biomarkers

eliminated via renal filtration, measurement error will decrease after correcting for SG,

CR, or urine output. However, the opposite trend was observed in the Ex-R study.

Specifically, measurement variability was the lowest both between- and within-

individuals for concentration-based measures. Correcting for CR, SG, and urine output

increased fold-range estimates in most instances. Furthermore, these correction

approaches increased between-person measurement variability more so than within-

person measurement variability, leading to elevated ICC estimates for corrected values.

While higher ICCs are beneficial from the standpoint of measurement reliability, they are

meant to reflect true exposure differences between individuals. In this study, it is

conceivable that the observed variability between subjects in urine output, CR excretion,

and SG levels, generally resulting from differences in hydration status, diet, body mass,

etc., are responsible for the heightened ICCs. Given these sources of measurement

variation, (reflecting physiological and behavioral factors and not pyrethroid exposure)

researchers are urged to use caution when selecting urinary biomarker-based exposure

metrics for 3-PBA as part of exposure and epidemiology studies in the future.

At the moment, it is unclear whether preformed metabolites of pyrethroid insecticides

(e.g., 3-PBA, cis-DCCA, and trans-DCCA) in media are substantially contributing to the

measured urinary pesticide metabolite levels in humans. Few studies have measured for

the concurrent levels of pyrethroids and their preformed metabolites in foods or other

environmental media (Starr et al., 2008; Chen et al., 2012; Trunnelle et al., 2014). Starr et

al. (2008) showed that permethrin, DCCA, and 3-PBA, were frequently detected (>67%)

in 85 vacuum dust samples collected at homes and childcare centers in NC and Ohio in

2000–2001. Chen et al. (2012) also reported detectable levels of several different

pyrethroids and 3-PBA in 23 different produce samples including blackberries,

blueberries, cherries, strawberries, and green onions in California in 2010. More recently,

Trunnelle et al. (2014) reported that permethrin and 3-PBA were both frequently detected

24

(>97%) in 79 kitchen floor wipe samples collected at homes in California in 2007–2009.

This above research suggests that humans can be exposed to measureable levels of

preformed pyrethroid metabolites in several different media (i.e., food, dust, and surface

wipes) at residences. Therefore, more research is necessary to determine if preformed

metabolites in media are overinflating measured urinary pyrethroid biomarker levels in

non-occupationally exposed adults. Future work is planned to quantify the distributions

of the target pyrethroids and their preformed metabolites in solid food, drinking water,

vacuum dust, and surface wipe samples collected at the participant’s residences in the Ex-

R study.

5. Conclusions

Based on the urinary biomonitoring data, the majority of the Ex-R adults were likely

temporally exposed to one or more pyrethroid insecticides at their homes during the six-

week monitoring period in NC in 2009–2011. For the frequently detected 3-PBA (>50%),

there was poor reproducibility of repeated measurements of this metabolite in the urine

samples of Ex-R adults regardless of sample type (bedtime, FMV, and 24-hr) and method

(concentration, SG-corrected, CR-corrected, and excretion rate) over a day, week, and six

weeks. Correcting for urine sample dilution, based on either SG, CR, or urine output, was

found to increase measurement variability between- and within-individuals. Given these

findings, scientists should use caution when selecting urinary biomarker-based exposure

metrics for 3-PBA in their future exposure and environmental epidemiology studies.

Acknowledgements

We sincerely thank the adult volunteers who participated in the Ex-R study. We also

thank Ronald Williams, James Starr, Denise McMillian, Kent Thomas, and Daniel Stout

II for their technical assistance during the study. We also gratefully acknowledge

WESTAT (Andrea Ware and Bethany Bradford) for their valuable assistance in

recruiting study participants. This work was funded in whole by the US EPA.

Disclaimer: “The US EPA through its Office of Research and Development has provided

administrative review of this article and approved it for publication.”

25

References

Andersen EM, Sobus JR, Strynar MJ, Pleil JD, Nakayama SF. Evaluating an alternative method for rapid urinary creatinine determination. J Toxicol Environ Health A. 2014:77: 1114–1123.

Baker SE, Olsson AO, Barr DB. Isotope dilution high-performance liquid chromatography-tandem mass spectrometry method for quantifying urinary metabolites of synthetic pyrethroid insecticides. Arch Environ Contam Toxicol. 2004:46:281–288.

Barr DB, Wilder LC, Caudill SP, Gonzalez AJ, Needham LL, Pirkle, J.L. Urinary creatinine concentrations in the US population: implications for urinary biologic monitoring measurements. Environ Health Perspect 2005;113:192–200.

Barr DB, Olsson AO, Wong LY, Udunka SO, Baker SE, Whitehead RD, et al. Urinary concentrations of metabolites of pyrethroid insecticides in the general US population: National Health and Nutrition Examination Survey 1999-2002. Environ Health Perspect 2010;118:742–748.

Berkowitz GS, Obel J, Deych E, Lapinski R, Godbold J, Liu Z, et al. Exposure to indoor pesticides during pregnancy in a multiethnic, urban cohort. Environ Health Perspect 2003;111:79–84.

Boeniger MF, Lowry LK, Rosenberg J. Interpretation of urine results used to assess chemical exposure with emphasis on creatinine adjustments: A review. Am Ind Hyg Assoc J 1993; 54: 615–627.

CDC (Centers for Disease Control and Prevention). Fourth national report on human exposure to environmental chemicals. 2009. [http://www.cdc.gov/exposurereport/ (5 May 2015)].

CDC (Centers for Disease Control and Prevention). Fourth national report on human exposure to environmental chemicals; Updated tables. 2015. [http://www.cdc.gov/biomonitoring/pdf/FourthReportUpdatedTables_Feb2015.pdf (11 May 2015)].

Chen L, Zhao T, Pan C, Ross JH, Krieger RL. Preformed biomarkers including dialkylphosphate (DAPs) in produce may confound biomonitoring in pesticide exposure and risk assessment. J Agric Food Chem 2012;60:9342–9351.

Christensen KL, Lorber M, Koch HM, Kolossa-Gehring M, Morgan MK. Population variability of phthalate metabolites and bisphenol a concentrations in spot urine samples versus 24-h or 48-h collections. J Expo Sci Environ Epidemiol 2012;22:632–640.

26

Chuang JC, Wilson NK. Multiresidue analysis of organophosphate and pyrethroid pesticides in duplicate-diet solid food by pressurized liquid extraction. J Environ Sci Heal B 2011; 46:41–40.

Dewailly E, Forde M, Robertson L, Kaddar N, Sidi EA, Cote S, et al. Evaluation of pyrethroid exposure in pregnant women from 10 Caribbean countries. Environ Int 2014; 63:201–201.

Elliott M. Properties and applications of pyrethroids. Environ Health Perspect 1976;14: 3–13.

Fleiss, J. The Design and Analysis of Clinical Experiments; John Wiley & Sons: New York, NY, USA, 1985.

Fortin MC, Bouchard M, Carrier G, Dumas P. Biological monitoring of exposure to pyrethrins and pyrethroids in a metropolitan population of the Providence of Quebec, Canada. Environ Res 2008;107:343–350.

Hwang YH, Bornschein RL, Grote J, Menrath W, Roda S. Urinary arsenic excretion as biomarker of arsenic exposure in children. Arch Environ Health 1997;52:139–147.

Julien R, Adamkiewicz G, Levy J, Bennett DB, Nishioka M, Spengler JD. Pesticide loadings of select organophosphate and pyrethroid pesticides in urban public housing. J Expo Sci Environ Epidemiol 2008;18:167–174.

Kuhn KH, Wieseler B, Leng G, Idel H. Toxicokinetics of pyrethroids in humans: Consequences for biological monitoring. Bull Environ Contam Toxicol 1999;62:101–108.

Lassen TH, Frederiksen H, Jensen TK, Peterssen JH, Main KM, Skakkebaek NE, et al. Temporal variability in urinary excretion of bisphenol A and seven other phenols in spot, morning, and 24-h urine samples. Environ Res 2013;126:64–170.

Leng G, Kuhn KH, Idel H. Biological monitoring of pyrethroids in blood and pyrethroid metabolites in urine; Applications and limitations. Sci Total Environ 1997;199:173–181.

Lewis RC, Cantonwine DE, Anzalota Del Toro LV, Calafat AM, Valentin-Blasini L, Davis MD, et al. Urinary biomarkers of exposure to insecticides, herbicides, and one insect repellent among pregnant women in Puerto Rico. Environ. Health 2014;13:97.

Mage DT, Allen RH, Gondy G, Smith W, Barr DB, Needham LL. Estimating pesticide dose from urinary pesticide concentration data by creatinine correction in the Third National Health and Nutritional Examination Survey. J Expo Sci Environ Epidemiol 2004;14:457–465.

27

Mage DT, Allen RH, Kodali A. Creatinine corrections for estimating children’s and adult’s pesticide intake doses in equilibrium with urinary pesticide and creatinine concentrations. J Expo Sci Environ Epidemiol 2008;18;360–368.

McKelvey W, Jacobson JB, Kass D, Barr DB, Davis M, Calafat AM, et al. Population-based biomonitoring of exposure to organophosphate and pyrethroid pesticides in New York City. Environ Health Perspect 2013;121:1349–1356.

Morgan MK, Sheldon LS, Croghan CW, Jones PA, Chuang JC, Wilson NK. An observational study of 127 preschool children at their homes and daycare centers in Ohio: Environmental pathways to cis- and trans-permethrin exposure. Environ Res 2007;104: 266–274.

Morgan MK, Sheldon LS, Thomas KW, Egeghy PP, Croghan CW, Jones PA, et al. Adult and children’s exposure to 2,4-D from multiple sources and pathways. J. Expo. Sci. Environ. Epidemiol. 2008;18;486–494.

Morgan, MK. Predictors of urinary levels of 2,4-dichlorophenoxyacetic acid, 3,5,6-trichloro-2-pyridinol, 3-phenoxybenzoic acid, and pentachlorophenol in 121 adults in Ohio. Int J Hyg Environ Health (ePub Ahead of Print, 7 April 2015).

NPIC (National Pesticide Information Center). Bifenthrin: General fact sheet. 2011. [http://npic.orst.edu/factsheets/bifgen.html (11 May 2015)].

NRC (National Research Council of the National Academies of Sciences). Human Biomonitoring for Environmental Chemicals. The National Academies Press, Washington DC. 2006. p 215.

Palmquist K, Salatas J, Fairbrother A. Chapter 11; Pyrethroid insecticides: Use, environmental fate, and ecotoxicology. In Insecticides - Advances in Integrated Pest Management. 2012. [http://www.intechopen.com/books/insecticides-advances-in-integrated-pest-management/pyrethroidinsecticides-use-environmental-fate-and-ecotoxicology (4 November 2014)].

Qi X, Zheng M, Wu C, Wang G, Feng C, Zhou Z. Urinary pyrethroid metabolites among pregnant women in an agricultural area of the Province of Jiangsu, China. Int J Hyg Environ Health 2012;215;487–495.

Rappaport SM and Kupper LL. Quantitative Exposure Assessment. Publisher: SM Rappaport El Cerrito, California. 2008. p 168.

Ratelle M, Cote J, Bouchard M. Toxicokinetics of permethrin biomarkers of exposure in orally exposed volunteers. Toxicol Lett 2015;232:369–375.

28

Riederer AM, Bartell SM, Barr DB, Ryan PB. Diet and nondiet predictors of urinary 3-phenoxybenzoic acid in NHANES 1999-2002. Environ Health Perspect 2008;116:1015–1022.

Rosner B. Fundamentals of biostatistics. 6th edition, Duxbury Press, Pacific Grove, CA, 2006.

Saillenfait AM, Ndiaye D, Sabate JP. Pyrethroids: Exposure and health effects – An update. Int J Hyg Envir Heal 2015;218:281–292.

Schettgen T, Koch H, Drexler H, Angerer J. New gas chromatographic-mass spectrometric method for the determination of urinary pyrethroid metabolites in environmental medicine. J Chromatogr B 2002;778:121–131.

Starr J, Graham S, Stout D, Andrews K, Nishioka M. Pyrethroid pesticides and their metabolites in vacuum cleaner dust collected from homes and day-care centers. Environ Res 2008;108:271–279.

Stout DM, Bradham KD, Egeghy PP, Jones PA, Croghan CW, Ashley PA, et al. American Healthy Homes Survey: A national study of residential pesticides measured from floor wipes. Environ Sci Technol 2009;43:4294–4300.

Tao L, Chen M, Collins E, Lu C. Simultaneous quantitation of seven pyrethroid metabolites in human urine by capillary gas chromatography–mass spectrometry. J Sep Sci 2013;36:773–780.

Trunnelle KJ, Bennett DH, Tulve NS, Clifton MS, Davis MD, Calafat AM, et al. Urinary Pyrethroid and Chlorpyrifos Metabolite Concentrations in Northern California Families and Their Relationship to Indoor Residential Insecticide Levels, Part of the Study of Use of Products and Exposure Related Behavior (SUPERB). Environ Sci Technol 2014; 48: 1931–1939.

Tulve NS, Egeghy PP, Fortmann RC, Whitaker DA, Nishioka MG, Naeher LP, et al. Multimedia measurements and activity patterns in an observational pilot study of nine young children. J Expo Sci Environ Epidemiol 2008;18:31–44.

Ueyama J, Kimata A, Kamijima M, Hamajima N, Ito Y, Suzuki K, et al. Urinary excretion of 3-phenoxybenzoic acid in middle-aged and elderly general population of Japan. Env Res 2009;109:175–180.

US EPA (United States Environmental Protection Agency). Pesticide fact sheet: Metofluthrin. 2006. [http://nepis.epa.gov (7 May 2015)].

US EPA (US Environmental Protection Agency). Scientific and Ethical Approaches for Observational Exposure Studies (SEAOS). 2008. EPA/600/R-08/062. [http://www.epa.gov/nerl/sots/SEAOES_doc20080707.pdf (25 May 2015)]

29

US EPA (United States Environmental Protection Agency). Office of Pesticide Programs: Pyrethrins/Pyrethroid Cumulative Risk Assessment. 2011. [http://www.regulations.gov/ #!documentDetail;D=EPA-HQ-OPP-2011-0746-0003 (3 November 2014)].

US EPA (United States Environmental Protection Agency). Pyrethroids and Pyrethrins. 2013. [http://www.epa.gov/oppsrrd1/reevaluation/pyrethroids-pyrethrins.html (19 March 2015)].

Verbovsek, T. A comparison of parameters below the limit of detection in geochemical analyses by substitution methods. Materials and Geoenvironment 2011;58;393–404.

Wielgomas B. Variability of urinary excretion of pyrethroid metabolites in seven persons over seven consecutive days – Implications for observational studies. Toxicol Lett 2013(a); 221:15–22.

Wielgomas B, Nahorski W, Czarnowski W. Urinary concentrations of pyrethroid metabolites in the convenience sample of an urban population of Northern Poland. Int J Hyg Envir Heal 2013(b);216:295–300.

Xue J, Zartarian VG, Tornero-Velez R, Tulve NS. EPA’s SHEDs-multimedia model: Children's cumulative pyrethroid exposure estimates and evaluation against NHANES biomarker data. Environ Int 2014;73:304–311.

Young HA, Meeker JD, Martenies SE, Figueroa ZI, Boyd-Barr D, Perry MJ. Environmental exposure to pyrethroids and sperm sex chromosome disomy: A cross sectional study. Environ Health 2013;12:111.

30

Supplemental Information

The adult participant’s collected their own solid food, drinking water, surface wipes,

and vacuum dust samples at their homes over the six-week monitoring period. Duplicate

amounts of solid foods were collected by the participants on days 1-2 of each sampling

week. Liquid food samples were not collected as several pyrethroid insecticides were

detected at low levels in beverages (i.e., juices, sodas, and milk) in a previous study

(Morgan et al., 2007). For each sampling day, solid food samples were collected during

three consecutive time periods (as described above). Food samples for each time period

were placed into a pre-labeled 31 x 31 cm re-sealable polyethylene bag (Uline Shipping Supply Specialists) and then into a larger 41 x 41 cm re-sealable

polyethylene bag for secondary containment.

Drinking water samples (> 500 mL) were collected in a 1 L polypropylene container

on the morning of day 3 of the last sampling week at home from the participant’s primary

source of drinking water (e.g., well or municipal). The source of drinking water (well-

unfiltered, well-filtered, city-unfiltered, city-filtered, or bottled) was recorded on the

sample lid.

Surface wipe samples were collected in two high traffic areas (e.g., sink, stove, or

refrigerator) of the kitchen on day 4 of each sampling week. A 100% pre-cleaned cotton

pad (Twillwipes 100 cm2 - 1 ply) wetted with 10 mL of isopropanol was used to wipe

(left to right) a 960 cm2 area (using an aluminum template) of the kitchen floor. The

cotton pad then was folded in half (soiled side in) and the entire surface area was wiped

(left to right) again. The pad was then placed into a pre-cleaned 60 mL amber glass jar

31

with lid. The participant repeated the same procedure in another high traffic area of their

kitchen floor.

Dust samples were collected from carpets and/or hard floors (i.e., hardwood, tile, and

vinyl) of the main activity areas of the participants’ homes using their own vacuum

cleaners on day 4 of the last sampling week. Before sampling, participants with bag-style

vacuum cleaners removed their existing bag and replaced it with a new identical one.

Individuals with canister-style vacuum cleaners removed their existing filter and replaced

it with a new identical one and removed the existing dust sweepings. After vacuuming

the main activity areas, the bag or the filter/contents of the canister (> 2 cups) were

placed into a 31 x 38 cm re-sealable, polyethylene bag.

All environmental samples were stored in < -20oC freezers at the US EPA laboratory

until analyses.

32